Multisensor, Device-Based Monitoring of Patients with Systolic Heart Failure

Introduction 

Many patients with heart failure due to systolic dysfunction have recurrent heart failure events (HFE) in spite of optimal drug and device therapy. This patient population is also at risk for malignant ventricular arrhythmia, and are candidates for ICD or CRT-D implantation. Impedance data derived from implanted devices had shown promise for monitoring patient volume status^1,2 in this patient population, but studies subsequently showed no reduction in mortality or HFE.^3,4 A new device-based system has been developed that uses multiple sensors instead of impedance alone.⁵ A case demonstrating the use of this system is presented.

Case description

A 72-year-old female with an ischemic cardiomyopathy, left bundle branch block (LBBB), and NYHA class III heart failure symptoms underwent implant of a CRT-D. The implanted device (VIGILANT, Boston Scientific) provides standard CRT-D functionality and also integrates data from multiple sensors (heart sounds, respiration, activity level, transthoracic impedance, and night heart rate) to derive a composite score (HeartLogic index, Boston Scientific; Figure 1). The patient was enrolled in remote monitoring of her standard device parameters and her HeartLogic index via the manufacturer’s remote service (LATITUDE NXT, Boston Scientific). Approximately 6 weeks after implant, the patient’s HeartLogic index crossed the threshold for alert and our device clinic was notified via LATITUDE of the increased index (Figure 2). The patient was seen in our heart failure clinic and had noticed mild abdominal bloating, but no change in her physical exam, weight, or functional status. No change was made in her medical regimen. One week later, the HeartLogic index had increased further, due to increased S3, increased S3/S1 ratio, increased respiratory rate, and increased night heart rate (Figure 3). The transthoracic impedance was unchanged. The patient was contacted by phone; she stated that although her weight was unchanged, she had noticed more abdominal bloating and mild dyspnea on exertion. Her diuretic regimen was increased, and arrangements were made for a repeat heart failure clinic visit. On return to the clinic, the patient reported feeling significantly better with resolution of her abdominal bloating and dyspnea. A report of her HeartLogic index the next day showed a reduction to an index value of 9 (Figure 4) primarily due to a reduced S3, S3/S1 ratio, and to a lesser extent, by a reduction in night heart rate. The patient has remained clinically stable on her modified medical regimen. 

Discussion

Many patients with systolic heart failure have continued symptoms and heart failure admissions in spite of optimal drug and device therapy. Attempts to assist in the management of systolic heart failure using single sensor, device-based impedance monitoring has failed to show a reduction in mortality or heart failure admissions.^3,4 A multisensor system has been developed and studied in patients with systolic dysfunction receiving a CRT-D in the MultiSENSE trial.⁵ Data from five sensors was collected from 974 patients with implanted devices and the patients were followed for the development of HFE. The data from 531 of the patients was analyzed and used to develop the composite index. The MultiSENSE study found that when a weighted average from 5 of the sensors yielded an alert status at an alert threshold of 16 and a recovery threshold of 6, there was an 82% sensitivity for subsequent HFE and an unexplained alert rate of 1.33 per patient year in the development set. The algorithm was then applied to the data from the remaining sequestered 443 patients. In this group of patients, the same index threshold had a 70% sensitivity and an unexplained alert rate of 1.47 per patient year. 

The commercially approved index is known as HeartLogic, and is available in ICDs and CRT-Ds from Boston Scientific. The system uses a weighted average of data from heart sound (S3, S1), respiratory (rate, tidal volume), transthoracic impedance, and night heart rate sensors to generate the HeartLogic index. After implant, the feature is nominally on. Data collection from the sensors begins the next day, and requires 30 to 37 days for initialization. Index values range from 0 to 100. A large change in sensor data in the direction of worsening heart failure leads to a higher HeartLogic index value. Sensor data is aggregated at the end of each day and a HeartLogic index value is derived. Based on a programmable HeartLogic Alert threshold value (nominally 16), an alert status is assigned if the patient’s HeartLogic index is equal to or greater than the programmed threshold. An alert is then delivered via LATITUDE NXT remote monitoring and available for review. Weekly alerts will continue until the HeartLogic index falls below a recovery threshold (nominally 6).

In the MultiSENSE trial, the median time from alert onset to HFE was 34 days. Eighty-nine percent of patients who had an HFE had alert onset at least 2 weeks prior to the event. Thus, the HeartLogic system provides an early warning of potential heart failure events in the near future. It can potentially be used as a supplement to standard care by identifying patients who need closer follow-up. In our case, the initial alert prompted an early clinic visit, but evaluation at that time did not warrant intervention. One week later, the HeartLogic index had increased further, and the patient developed mild symptoms. Adjustment of the patient’s diuretic led to resolution of her symptoms and return of the HeartLogic index to non-alert value. 

Patients in the MultiSENSE trial had a 10-fold increased risk of HFE if their HeartLogic was in alert status. Preliminary data indicate that the use of NT-proBNP and the HeartLogic index together appears to further risk stratify patients.⁶ Patients with an NT-proBNP level of ≥1000 and in HeartLogic Alert at an alert threshold of 16 were 50 times more likely to have a HFE compared to patients with an NT-proBNP level of <1000 and out of HeartLogic Alert status at 16. Clinicians managing heart failure patients may want to consider using baseline NT-proBNP levels and HeartLogic monitoring to identify patients at particularly high risk for subsequent events for closer follow-up. No data is currently available that indicates management decisions utilizing the HeartLogic index will reduce heart failure events. However, a prospective, multicenter trial (MANAGE-HF) is currently enrolling to study this issue.

Conclusion

The HeartLogic feature of this patient’s CRT-D was successfully used to supplement our clinical judgement and standard of care by our heart failure clinic staff. A randomized trial is underway to determine whether use of this feature to guide therapy can reduce heart failure admissions. 

Disclosure: Dr. Rials was a co-investigator in the MultiSENSE trial, and has received speaking honoria from Boston Scientific.

This article was published with support from Boston Scientific.

References

1. Yu CM, Wang L, Chau E, et al. Intrathoracic impedance monitoring in patients with heart failure: correlation with fluid status and feasibility of early warning preceding hospitalization. Circulation. 2005;112:841-848.
2. Conraads VM, Tavazzi L, Santini M, et al. Sensitivity and positive predictive value of implantable intrathoracic impedance monitoring as a predictor of heart failure hospitalizations: the SENSE-HF trial. Eur Heart J. 2011;32:2266-2273.
3. Van Veldhuisen DJ, Braunschweig F, Conraads V, et al, DOT-HF Investigators. Intrathoracic impedance monitoring, audible patient alerts, and outcome in patients with heart failure. Circulation. 2011;124:1719-1726.
4. Luthje L, Vollmann D, Seegers J, Sohns C, Hasenfuß G, Zabel M. A randomized study of remote monitoring and fluid monitoring for the management of patients with implanted cardiac arrhythmia devices. Europace. 2015;17:1276-1281.
5. Boehmer J, Hariharan R, Devecchi FG, et al. A Multisensor algorithm predicts heart failure events in patients with implanted devices: results from the MultiSENSE Study. JACC Heart Fail. 2017;5:216-225.
6. Gardner RS, Singh JP, Stancak B, et al. HeartLogic Multisensor algorithm identifies patients during periods of significantly increased risk of heart failure events: results from the MultiSENSE study. Circ Heart Fail. 2018;11:e004669.